| Beyond 5th generation(B5G)mobile communications are expected to provide ultra-high spectral efficiency and transmission rates,ultra-reliable and low-latency communications,and large-scale Internet of Things ecosystems to meet the growing requirements of mobile communications.Also,B5 G mobile communication systems require sufficient flexibility and intelligent capabilities such as network self-awareness and self-adjustment to cope with the rapid and unexpected changes in future mobile communications.There are several key technologies for B5 G mobile communications such as large-scale antenna arrays,dense networks,millimeter-wave access,new waveform multiplexing,channel coding/decoding,machine learning,etc.In order to improve the data rate,this dissertation investigates adaptive transmission theory and method for B5 G mobile communications.The main contributions are listed as follows.Firstly,the adaptive modulation and power allocation are investigated for multi-user multicarrier system.The main work includes: 1)To maximize the data rate,adaptive modulation is applied for each subcarrier,where the M-quadrature amplitude modulation order is designed to meet a preset bit error rate(BER)requirement;2)Combining the modulation order,power allocation scheme for each subcarrier is designed to maximize the data rate with the transmission power constraint;3)The effect of carrier frequency offset(CFO)is investigated and analyzed,and it is proved that different CFOs will cause different interference variances to adjacent users,which results in performance degradation in universal filtered multi-carrier(UFMC)systems.Simulation results show that the proposed adaptive modulation and power allocation algorithm can meet the bit error rate performance requirements with the maximum data rate,and the simulation results are corresponding to the theoretical analysis that CFOs lead to performance degradation in UFMC systems.Secondly,the adaptive filter design problems are investigated for multi-user new waveform UFMC systems with CFOs.The main work includes: 1)The finite impulse response filter based on the weighted Chebyshev approximation is employed for adaptive filter scheme;2)The passband and stopband parameters of the filter are designed according to the bandwidth of different users and the guard interval between adjacent users;3)With the BER constraint,the stopband ripple,as the main parameter of the filter,is designed according to the different CFOs of different users.Simulation results show that the proposed adaptive filter configuration algorithm can dramatically eliminate the interference caused by different CFOs,and achieve better BER performance and a higher data rate than the conventional scheme.Thirdly,the beam-domain adaptive transmission scheme is investigated for hybrid architecture massive multiple input multiple output(MIMO)systems.The main work includes: 1)A simple two-dimensional direction of arrival(DOA)estimation method is designed by using the limited radio frequency chains in hybrid architecture massive MIMO systems;2)A lowcomplexity channel estimation method is proposed,which utilizes the beam steering vectors achieved from the DOA estimation and beam gains estimated by low-overhead pilots;3)Based on the estimated beam information,a purely analog precoding/combining strategy is designed,and the optimal power allocation among multiple beams is derived to maximize data rate;4)A novel adaptive beam management scheme is proposed to quickly determine the optimal beam,which greatly reduces the time cost of beam selection and beam access.Simulation results show that the proposed beam domain signal processing schemes have reduced complexity,high channel estimation accuracy and data rate.Fourthly,high data rate and performance analysis problems are investigated for deep learning(DL)based communication systems,where adaptive theory is employed.The main work includes: 1)With a mean squared error constraint,an adaptive transmission is designed to select the transmission vectors for maximizing the data rate under different channel conditions;2)A new generalized data representation(GDR)scheme is proposed to improve the data rate of DLbased communication systems;3)The effect of training signal-to-noise ratio(SNR)is analyzed for these DL-based communication systems.Numerical results show that the proposed adaptive transmission and GDR schemes can achieve higher data rate and lower training complexity when they have comparable block error rate(BLER)performance compared to the conventional one-hot vector scheme.Theoretical analysis and simulation results show that high training SNR or training SNR set can attain robust BLER performance under different channel conditions. |
PDF Full Text Request